DataDomain NFS Gateway

[captainflannel]

We recently purchased a DD2500 and are using a Linux Virtual Machine for NFS mount access to our DD. Currently the Linux NFS Gateway VM has 4vCPU and 20GB of memory, our veeam proxys, 4vCPU and 6GB RAM. We have up to 3 virtual proxys, with 4 concurrent tasks each for any given backup job. Backup jobs are setup with no compression and no deduplication, we have everything processed on the DD.
During VEEAM backup jobs our Linux NFS Gateway gets significant memory utilization, 98%. I've continued to add memory to the VM and it continues to be consumed. How can I determine an appropriate memory size for optimal performance? I was assuming 2GB per concurrent task, but I'm not sure at this point.

[Vitaliy S.]

Hello,

What processes consume the allocated memory on this box, can you please double-check? How many tasks do you have configured on the repository?

Thank you!

[Gostev]

Physical limitation is 3GB per job (due to 32-bit data mover process), so you should not be seeing more than that under any circumstances.

[captainflannel]

Regarding specific process consuming memory, there is only the various veemagent/veeam-soap processes. When I sort by memory there is no single process which takes over 700mb.

Regarding repository concurrent task limits, at the moment we do not have that enabled.

I'll recalculate to 3GB per task and update.